Committees of deep feedforward networks trained with few data


Deep convolutional neural networks are known to give good results on image classification tasks. In this paper we present a method to improve the classification result by combining multiple such networks in a committee. We adopt the STL-10 dataset which has very few training examples and show that our method can achieve results that are better than the state of the art. The networks are trained layer-wise and no backpropagation is used. We also explore the effects of dataset augmentation by mirroring, rotation, and scaling.
Original languageEnglish
Number of pages7
Publication statusPublished - 23.06.2014


Dive into the research topics of 'Committees of deep feedforward networks trained with few data'. Together they form a unique fingerprint.

Cite this